Landslide susceptibility mapping using fuzzy logic and UAV imagery in GIS framework: a case study of Baling, Malaysia

Landslides threaten sustainable development in tropical regions like Malaysia, where climate change intensifies rainfall-induced slope failures. This study presents a low-impact approach combining unmanned aerial vehicles (UAVs), fuzzy logic, and GIS to map landslide susceptibility in Baling, Kedah, which is a region experiencing recurrent slope instability. The primary objectives are to (1) quantify the influence of slope, rainfall, soil type, and land cover on landslide risks, (2) evaluate fuzzy logic's effectiveness in handling geospatial uncertainties, and (3) provide actionable insights for policymakers to support Sustainable Development Goals (SDGs) 11 and 15. High-resolution UAV orthophotos (2cm) and GIS datasets were processed using fuzzy logic, with membership functions classifying slope angles (low: 0°–15°; moderate: 15°–30°; high: >30°) and rule-based analysis. Validation against historical landslide data yielded strong accuracy (MAE = 0.12; RMSE = 0.18). Results identified the northwest region as high-risk (susceptibility index >0.7) due to steep slopes (>30°) and clayey soils, while the eastern sector exhibited low risk (index <0.3) with gentler slopes (<15⁰). The model achieved 85% accuracy, outperforming conventional methods. This study contributes a scalable, cost-effective framework for landslide risk assessment, particularly valuable for developing regions. It supports sustainable development goals (SDGs) by providing a low-cost, high-precision landslide assessment tool deployable in resource-limited regions.